Faculty Advisor

Yeesock Kim

Faculty Advisor

Tahar El-Korchi

Faculty Advisor

Leonard Albano

Identifier

etd-041812-134324

Abstract

With the increased deterioration of infrastructure in this country, it has become important to find ways to maintain the strength and integrity of a structure over its design life. Being able to control the amount a structure displaces or vibrates during a seismic event, as well as being able to model this nonlinear behavior, provides a new challenge for structural engineers. This research proposes a wavelet-based adaptive neuro- fuzzy inference system for use in system identification and structural control of civil engineering structures. This algorithm combines aspects of fuzzy logic theory, neural networks, and wavelet transforms to create a new system that effectively reduces the number of sensors needed in a structure to capture its seismic response and the amount of computation time needed to model its nonlinear behavior. The algorithm has been tested for structural control using a three-story building equipped with a magnetorheological damper for system identification, an eight-story building, and a benchmark highway bridge. Each of these examples has been tested using a variety of earthquakes, including the El-Centro, Kobe, Hachinohe, Northridge, and other seismic events.

Publisher

Worcester Polytechnic Institute

Degree Name

MS

Department

Civil & Environmental Engineering

Project Type

Thesis

Date Accepted

2012-04-18

Award

Sigma Xi Graduate Research Award for Outstanding Doctoral Dissertation (2015)

Accessibility

Unrestricted

Subjects

system identification, structural control, neural network, fuzzy logic, wavelet transform, earthquake

Share

COinS